Mohanad's "Blog" πŸ€“πŸ‘‹

AI in Arabic - The Vision

AI in Arabic was created in October 2019, with the first article published on October 11, 2019 and was officially launched on the first of April 2020. The team has been growing since then and reached twenty-three members in between bloggers, reviewers and two managers. In order for everyone to stay on the same page, it is critical to have a clear vision and strategy, which will help in having a crystal clear image of how the project will look like a few years from now.

In this document, we will discuss the vision and the strategy needed to achieve this vision. It is important to have a realistic and achievable strategy in order to accomplish the goal of this project.

A valid question to ask, why this blog post is written in English while the project is about Arabic. The answer is because I don’t have an Arabic keyboard πŸ˜›.

AI in Arabic Mission and Vision πŸš€

The mission of AI in Arabic is to provide the Arabic community with the needed content and tools in the Arabic language in order to accelerate and advance the research and the achievements in this domain.

The vision of this project can be divided into three levels, I like to represent it with circles as shown in the figure 1 below. The inner small circle can be represented by increasing the awareness of artificial intelligence by writing and translating articles into the Arabic language. Then the next circle is represented by enabling the community to start producing content and working on projects by providing tools and means like a dictionary and libraries to start using and utilizing AI’s algorithms. The outer third circle should empower entrepreneurs by fostering and creating labs and connecting them together by providing the networking and the funding as well as the technical support needed to leverage and build solutions in the AI domain.

The Strategy πŸ—ΊοΈ

The strategy can be divided into three phases. Phase 1 increases the awareness of artificial intelligence, phase 2 - empower the community by creating tools and libraries, phase 3 - empower the community to start making solutions and start ups in the AI domain.

Phase 1 - Stabilize Publishing Articles ✍️

We are in the inner circle, where we have a blog where we translate articles from English to Arabic. In addition, we standardized the workflow of writing articles which helps in increasing and controlling the quality of the published articles as well as making the process scalable.

But, the process is still not stable. For example, the quantity of published articles is not stable and fluctuates from month to month with 1 to 2 articles per month on average.

Before we can move to the next phase in the vision, we need to stabilize the process of writing and publishing articles -- meaning we should have a solid base before we can start creating tools or even building a dictionary.

The current process of writing articles can be improved by the following points: 1. Create and iterate on a solid path of articles -- currently we do have a preliminary path per category (e.g, machine learning path) but we need to make them more visible to the whole team as well as have a selected articles and topic to work on. Meaning, create a bank of tasks based on each path and have all the needed resources ready per task.

  1. To maintain a regular number of published articles, we should always have a minimum number of active bloggers. For example, if the goal is to write at least 4 articles per month, we will need at least 4 active bloggers. We should remember the life cycle to active bloggers is less than two articles, meaning we should always be looking for new candidates.

  2. Create and provide the needed documentation to make writing articles as easy as possible. Currently, we do have few documents which explain how to write articles or how to publish on Wopdpress, but many of them are outdated or not visible to bloggers.

  3. Have a unified communication tool (for example, Slack or Element)

  4. Generate and utilize analytics. Which is critical to know how many visitors are using the platform. (Google Analytics is already set up for the AI in Arabic Blog). In addition, create metrics to help get an idea of how we are doing.

  5. Engage with the community. For example, as of now, the only engagement we have is to list the published articles on the Facebook page. We will need to increase the engagement with the community.

  6. Collaborate with other platforms.

By stabilizing the publishing process and iterating on it based on the feedback we get from the defined metrics we can define a goal of when we can move to the next goal (circle).

Phase 2 - Empowering the Community 🧰

In this phase, which is represented by the middle circle of the vision above, AI in Arabic will focus on providing the community with tools, libraries and a dictionary, which will help and enable students, researchers as well as entrepreneurs to build solutions as well as writing and conducting research in the AI domain.

In this phase, AI in Arabic should focus on the following:

English <> Arabic Dictionary. A specialized dictionary for terms in the domain of AI. In order for the dictionary to be usable and trust worthy by the community, we should take the followung into consideration:

  1. Have a rich translation. Meaning, every translated term MUST have a description of the translation, as well as providing examples.

  2. Standardize the process of translating a term. For example, how to translate a term, who should approve the translation, how many people should approve a translation.

  3. Preferably, engage accredited organization within the translation (for example, universities)

  4. Introduce the dictionary within the academic community so more students and researches can use it

  5. Create tools to make the translation easier

  6. Create a standalone website only for the dictionary with an easy format to search and browse

  7. Libraries and APIs: develop and create libraries and APIs that can be used by the community members. In addition, we can work on collecting and creating a database of training sets which can also be utilized by community members. The APIs can be free or paid.

  8. Professional Training: this can be anything from training on frameworks like Tensorflow or even teaching on topics like computer vision. This can also be a mix between free and paid.

The outcome of this phase will be: AI in Arabic will be trusted and well known by the academic as well as the professional community AI in Arabic, unlike in phase 1, will be able to make some revenue by monetizing the

Libraries and the training: develop and create libraries and APIs that can be used by the community members. In addition, we can work on collecting and creating a database of training sets which can also be utilized by community members. The APIs can be free or paid.

Professional Training: this can be anything from training on frameworks like Tensorflow or even teaching on topics like computer vision. This can also be a mix between free and paid.

The outcome of this phase will be:

  • AI in Arabic will be trusted and well known by the academic as well as the professional community

  • AI in Arabic, unlike in phase 1, will be able to make some revenue by monetizing the libraries and the training.

Phase 3 - Empowering Entrepreneurs πŸ§‘β€πŸš€

This will be the final phase of AI in Arabic. In this phase, both the community as well as AI in Arabic, will become more mature to start fostering and creating technical solutions by leveraging artificial intelligence techniques to solve real world problems.

AI in Arabic will serve in three points:

  1. Make our own solutions aka startup(s)

  2. Fostering and investing in other startups by providing the technical as well as the financial support

  3. Create a strong network and build bridges in between the industry demand and the academia supply which helps in closing the gap between the supply and demand.

Conclusion

In conclusion, the vision is divided into three phases, where the first phase is represented by translating and publishing articles, phase 2 by creating tools, libraries and a dictionary to help the community to leverage AI techniques, and the last phase will foster and invest in solutions in the field of artificial intelligence.